Faster Is Better
Implications for Speed-Intensive Gait Training After Stroke
Anouk Lamontagne, PhD, PT; Joyce Fung, PhD, PT
Background and Purpose—The instantaneous adaptations to speed and load changes during overground locomotion have
major implications for mobility after stroke. We examined the extent to which stroke subjects could increase their
overground walking speed with respect to speed and unloading changes.
Methods—Twelve subjects with a unilateral stroke were evaluated while walking overground full weight bearing (FWB)
or with body weight support (BWS) at preferred or fast speed. On the basis of their preferred walking speed, subjects
were classified as high (ⱖ45 cm/s) or low functioning (⬍45 cm/s). Gait speed, temporal distance factors (TDFs), as well
as movements and muscle activation of the lower limbs were measured and compared across the conditions.
Results—FWB-Fast condition induced marked (165%) increment in gait speed in all subjects. BWS at preferred speed
induced faster speeds in low- but not the high-functioning subjects, whereas combined BWS and fast walking yielded
further speed increments in the high-functioning subjects. Fast walking was associated with bilateral increases in joint
excursion and muscle activation, as well as improved symmetry in some TDFs. BWS favored a hip flexion strategy in
early swing while decreasing limb circumduction.
Conclusions—This study shows that stroke subjects can increase substantially their walking speed without deleterious
effects. Fast walking induces marked speed-related improvements in body and limb kinematics and muscle activation
patterns. BWS during overground walking also increases gait speed, but to a lesser extent and only in low-functioning
subjects. The combination of BWS with fast speed produces the greatest increments in walking speed in all subjects.
(Stroke. 2004;35:000-000.)
Key Words: exercise 䡲 hemiplegia 䡲 locomotion 䡲 rehabilitation
W
alking after stroke is characterized by slow gait speed,1
poor endurance,2 and changes in the quality3 and
adaptability4 of walking pattern. Among rehabilitation paradigms developed to improve mobility in stroke subjects, the
most popular is treadmill training, alone5,6 or combined with
partial body weight support (BWS).7–9 Such training paradigms, in which subjects are trained at their preferred speed,
lead to improvements in gait speed that are modestly superior
to those obtained by conventional gait training.5,6 More
recently, new training paradigms incorporating fast walking
and intensive training regimens were proposed.10 –12 These
paradigms, incorporating principles of sports training such as
task-specificity, repetition, and training intensity, lead to
dramatic improvements in walking speed.12 However, little is
known about the stroke subjects’ capability of adapting to the
demands of speed and load, depending on their initial
functional performance.
The majority of training regimens involve treadmill walking (TW) with BWS. However, TW differs from overground
locomotion in several aspects,13,14 and the gains in walking
speed achieved during treadmill training by hemiparetic
subjects are not completely transferred to overground locomotion.6,9,12 A new system developed in our laboratory
allows subjects to walk overground while the lower extremities are partially unloaded.15,16 In the present study, the
instantaneous adaptations to fast walking and to BWS were
examined during overground locomotion in a group of
subjects with recent stroke. The purpose of this study was to
investigate the extent to which stroke subjects can increase
their speed and modify their gait pattern during (1) fast
walking, (2) BWS, and (3) combined fast walking and BWS.
Methods
Subjects
Twelve subjects were recruited from the Jewish Rehabilitation
Hospital with informed consent. Subjects (with characteristics listed
in the Table) experienced a cerebrovascular accident ⬍1 year
previously in the territory of the middle (n⫽11) or posterior (n⫽1)
cerebral artery, either of ischemic (n⫽10) or of hemorrhagic (n⫽2)
origin. Inclusion criteria were: first time unilateral stroke; ability to
walk 5 consecutive steps, either with or without assistance or
walking aid; and gait speed slower than 75 cm/s. Exclusion criteria
included brain stem or cerebellar lesion confirmed by computerized
Received August 4, 2004; accepted August 10, 2004.
From the Jewish Rehabilitation Hospital, Centre for Interdisciplinary Research in Rehabilitation; and the School of Physical and Occupational Therapy,
McGill University, Montreal, Canada.
Correspondence to Dr Anouk Lamontagne, JRH Research Centre, 3205 Place Alton-Goldbloom, Laval (Quebec), Canada, H7V 1R2. E-mail
anouk.lamontagne@mcgill.ca
© 2004 American Heart Association, Inc.
Stroke is available at http://www.strokeaha.org
DOI: 10.1161/01.STR.0000144685.88760.d7
1
2
Stroke
November 2004
Subject Characteristics
Subject Gender
Age
(years)
Paretic
Side
Time Since
Stroke (days)
Gait Speed
(cm/s)
Functional
Level*
1
Male
88
Right
69
9
Low
2
Male
61
Left
46
13
Low
3
Male
70
Left
39
32
Low
4
Female
52
Left
47
34
Low
5
Male
54
Left
290
37
Low
6
Male
78
Left
31
42
Low
7
Male
76
Right
128
43
Low
8
Male
64
Right
55
45
High
9
Male
69
Right
47
45
High
10
Female
74
Left
36
53
High
11
Female
73
Left
39
63
High
12
Male
82
Right
41
73
High
Mean
—
70
—
72
41
—
SD
—
11
—
73
18
—
*Based on their gait speed, subjects were classified as low- (⬍45 cm/s) or high- (ⱖ45 cm/s)
functioning.
tomography and any pre-existing conditions that interfere with
locomotion. A previous study by Perry et al17 shows that walking
speed of ⱕ0.4⫾0.18 m/s restricts a stroke individual’s capacity for
community ambulation. Thus, subjects were further classified (Table), on the basis of their preferred walking speed, as low (⬍45 cm/s)
or high (ⱖ45 cm/s) functioning. In the low-functioning group, 5
subjects had started to walk ⬍1 week previously.
tendinosus [ST], and rectus femoris [RF]) was recorded at 1080 Hz
using an 8-channel EMG system (TelEMG; BTS). After preamplification and band-pass filtering (10 to 400 Hz), EMG signals were
rectified and smoothed at 20 Hz, whereas kinematic data were
low-pass filtered at 8 Hz for offline analysis.
Experimental Design
The primary outcome measure of this study was gait speed. Secondary outcome measures included temporal distance factors (TDFs),
sagittal plane segmental angular excursions, CoM, and toe trajectory
along the mediolateral (ML) axis, as well as EMG amplitude. TDFs
retained for analysis were step length and cycle duration, as well as
the percentage of time spent in stance, single, and double support
periods. Foot strikes and foot-off events were determined using toe
and heel marker displacements along the anteroposterior (AP) axis.
All data were normalized to 100% of the gait cycle duration.
Walking speed was calculated as the average velocity of the body
CoM along the direction of walking for each gait cycle. EMG
integrals were computed for functionally relevant time windows of
the gait cycle (%): MG activation at push-off [30%:70%], TA
The evaluation consisted of 4 walking conditions presented randomly: walking full weight bearing (FWB) at preferred (FWB-Pref)
and maximal speed (FWB-Fast), and walking with BWS at preferred
speed (BWS-Pref) and maximal speed (BWS-Fast). Two walking
trials were assessed for each of the 4 conditions.
Walking Assessment
Subjects were evaluated on a 10-m walkway while walking in a body
harness suspended overhead by a pressurized weight support system.
This unique system provides constant BWS throughout the gait cycle
while not preventing the center of mass (CoM) of the body to travel
naturally up and down during the single- and double-limb support
periods. Subjects were instructed to walk at preferred speed and as
fast as possible with FWB or with BWS. To further encourage the
subjects during the fast walking conditions, they were told to hurry
as much as possible, as if they were trying to catch a bus, and
reinforced with verbal encouragement and cheering throughout the
walking trial. They were not discouraged from running if they could.
Level of BWS was set at 30% of body weight for most subjects
because such a level of unloading has been shown previously to yield
positive results for hemiparetic subjects.8,9,18 However, for 2 subjects
of the low-functioning group, 50% of BWS was provided because
this was preferred by the individual and allowed a better swingthrough with the paretic limb. Subjects wore a harness at all times,
and rest periods were given as needed. The walking speed or cadence
during different loading conditions was not dictated by the use of a
metronome or any artificial device. Instead, the different walking
conditions were assessed as they would be presented to the patient in
the clinical setting, and speed was our primary outcome of interest.
A Vicon 512 6-camera motion analysis system (Oxford Metrics)
was used to acquire body movements at 120 Hz. A 9-segment model
was reconstructed on the basis of a 23-marker set (PlugInGait;
Vicon) comprising the lower body segments, pelvis, trunk, and head.
Electromyographic (EMG) from paretic and nonparetic lower limb
muscles (medial gastrocnemius [MG], tibialis anterior [TA], semi-
Data Analysis
Figure 1. Average change in walking speed across the different
conditions with reference to the preferred speed condition at
FWB. The walking conditions illustrated are: FWB-Fast, BWSPref, and BWS-Fast.
Lamontagne and Fung
Fast Walking Paradigm in Stroke
3
Figure 2. Average walking speed of the
stroke subjects (A) and respective values
of cycle duration (B) and step length (C)
for the paretic and nonparetic limbs
while walking FWB or with BWS at preferred or fast speed. Values illustrated
represent the mean ⫾95% CI. For all
figures, only significant main and interaction effects of load (full weight vs BWS),
speed (preferred vs fast), or side (paretic
vs nonparetic) are indicated on each
graph.
activation at toe-off [60%:80%], ST activation in early stance
[0%:30%], and RF hip flexion burst at toe-off [60%:80%]. All
outcome measures were first calculated separately for every gait
cycle (n⫽8 to 12 cycles) and later averaged for every subject. The
first and last 2 gait cycles of each walking trial were always excluded
from the analysis to avoid acceleration and deceleration phases. Two
subjects ran for a few steps in one of the FWB-fast condition trials,
but those gait cycles were not included in the analysis.
Figure 3. Percentage of time spent in
the stance phase, as well as the percentage of time spent in total double-support
(B) and single-support (C) subphases on
the paretic and nonparetic sides while
walking FWB or with BWS at preferred
or fast speed. Values illustrated represent the mean ⫾95% CI.
4
Stroke
November 2004
Statistical Analysis
The effect of load (FWB versus BWS) and speed (Pref versus Fast)
on gait speed was analyzed using a 2-way ANOVA. The effects of
load, speed, and limb (paretic versus nonparetic) on all other
kinematic, TDF, and EMG variables were determined by using
3-way ANOVAs (Statistica 6.0). A P value of 0.05 was accepted as
significant after Bonferroni adjustments. When significant main or
interaction effects were present, post hoc comparisons were made by
Tukey tests.
Results
The average change in gait speed attributable to fast walking
without BWS was much greater in the low-functioning subjects
(210%) compared with the high-functioning group (103%),
although similar absolute speed increments were displayed in
both groups (␦Low Level⫽55 cm/s versus ␦High Level⫽56 cm/s; Figure
1). BWS alone induced faster speeds in the low-functioning
(␦⫽21 cm/s or 69%) but not the high-functioning (␦⫽⫺1 cm/s
or ⫺2%) group. The combination of BWS and fast walking
induced greatest increments in gait speed, revealing an interaction effect attributable to speed and BWS (␦Low Level⫽72 cm/s,
258% versus ␦High Level⫽108 cm/s, 95%; Figure 1), depending on
the functional level.
Increase in gait speed in the fast walking conditions was
accompanied by shorter cycle durations and larger step
lengths, regardless of BWS (Figure 2). Figure 3 shows that
fast walking induces a lesser proportion of stance duration,
along with shorter double-support and longer single-support
durations. Fast walking improved symmetry in double- and
single-support proportions between the paretic and nonparetic
sides, whereas BWS reduced stance proportion without significant improvements in symmetry.
When walking at faster speeds, subjects displayed bilateral
and symmetrical increases in hip and knee excursions. The
latter was explained by greater knee flexion at weight
acceptance and during swing, whereas larger hip excursions
were associated with increased extension in late stance and
flexion in late swing (P⫽0.001 to 0.04). Walking with
BWS-Pref or BWS-Fast induced larger hip excursions on the
paretic and nonparetic sides because of larger (P⬍0.01)
late-stance hip extension and swing-phase hip flexion. Fast
walking or BWS did not improve symmetry of the lower limb
movements. The occurrence of undesirable movement patterns such as increased limb circumduction or large side-toside displacement of body CoM was also examined. Fast
walking slightly increased the ML deviation of the paretic
and nonparetic foot trajectories, whereas BWS had the
opposite effect of slightly decreasing the ML deviation
(Figure 4). Side-to-side movements of the body CoM were
reduced at faster walking speeds (P⬍0.001; data not shown),
whereas BWS did not induce any significant effects
(P⬎0.05).
Fast walking required larger muscle activation in all
recorded muscles (Figure 5) on the paretic and nonparetic
sides. BWS did not induce any changes, except for RF, for
which activation increased at faster speed, especially on the
paretic side.
Discussion
Faster Is Better
Stroke subjects can increase their overground walking speed
2 to 3 times beyond comfortable levels given proper instruc-
Figure 4. The top traces (A) represent one example of the ML
displacement of the foot (x axis) as a function of stride length (y
axis) for the paretic (left) and nonparetic (right) limbs across the
different walking conditions: FWB-Pref, FWB-Fast, BWS-Pref,
and BWS-Fast. Bottom graphs (B) represent the mean (⫾95%
CI) ML excursion of the foot displacements during the same
walking conditions.
tions and a safe environment. Such increase is attributable to
larger step lengths and shorter cycle durations on the paretic
and nonparetic sides, regardless of the functioning level.
Muscle strength, more specifically, ankle push-off during the
stance phase and hip flexion pull-off in early swing,19,20 are
known to be major determinants of walking speed after
stroke. Slow walking after stroke may be a behavioral
adaptation to poor endurance, poor balance, and the individual’s perceived limits of stability. Use of a safety harness may
have unmasked the latent locomotor behavior and its capability to be driven fast. Other common factors after stroke,
such as altered perception of self-motion,21 and cognitive or
attention deficits,22 may also play a role.
Fast walking improves the kinematics and muscle activation patterns of hemiparetic gait, which are consistent with
the speed-dependent changes reported in healthy subjects.23
In contrast to fast TW,24 symmetry of some TDFs also
improved with overground fast walking, suggesting that
different strategies may be used in adapting to different
modes of locomotion (treadmill versus overground). Although these improvements provide a strong rationale for
training, the question of fatigability, energy cost, and pres-
Lamontagne and Fung
Fast Walking Paradigm in Stroke
5
Figure 5. Activation amplitude of the
MG, TA, ST, and RF on the paretic and
nonparetic side while walking FWB or
with BWS at preferred or fast speed.
Titles of each graph indicate the interpreted muscle function on the basis of
the time window during which activation
amplitude was measured during the gait
cycle. Mean values and respective 95%
CIs are illustrated.
ence of comorbidities may be raised. Paretic gait is known to
be associated with a high-energy demand compared with
normal gait,25 and stroke subjects show poor endurance.2
However, despite higher heart rates and larger muscle activation levels at faster speeds, stroke subjects’ overall energy
and cardiac costs correlate negatively with gait speed,24
suggesting greater efficiency at faster walking speeds. In fact,
when walking faster, stroke subjects approach the bandwidth
[1.2 to 1.8 m/s] of walking speed at which optimal upper body
segment coordination26 and energy consumption27 are observed. This greater efficiency may be attributable to more
appropriate timing of lower limb muscles,24 improved movement coordination, and possibly facilitation of intralimb and
interlimb energy transfers. Despite these advantages of fast
walking, its use may not be recommended for patients with
comorbidities such as heart disease or exercise-exacerbated
muscle/joint pain, unless close monitoring of vital signs,
exertion, or pain level allows for safe and comfortable use.
BWS Is Beneficial to Low-Functioning Subjects
In the present study, only subjects walking at an initial speed
⬍45 cm/s increased their comfortable speed when walking
with BWS. High-functioning subjects did not increase their
comfortable speed with BWS but did so when walking at
maximal pace. These results suggest that BWS may be
preferentially beneficial to low-functioning subjects, whereas
high-functioning subjects may benefit from the combined use
of BWS and fast walking. As observed in supported TW,18
BWS did not significantly improve symmetry in TDFs.
However, in contrast to what was reported for healthy28 and
stroke subjects18 during supported treadmill ambulation, no
reduction in antigravitary muscle activation was observed in
the present study. The lower demand on antigravitary muscles
induced by the unloading of the lower limbs may be counteracted by a higher demand because of the increase in
walking speed, which outlines a possible advantage of
overground BWS training over weight-supported TW. The
increase in RF activation at toe-off, especially on the
paretic side, suggests the use of a hip flexor strategy. This
strategy, along with the reduced limb circumduction observed with BWS, is consistent with the patients’ selfreport of increased ease in limb swing with BWS. This
improvement may be explained by the more upright
position of the trunk with BWS,29 which facilitates terminal stance hip extension and, in return, recruitment of the
hip flexors in early swing. Apart from facilitating stepping,
BWS during overground walking also increases the relative loading of the paretic limb and assists balance.29 These
effects on stepping, weight bearing, and balance, combined
with the speed increments experienced by low-functioning
stroke subjects in the present study, suggest overground BWS to
be useful for locomotor training early after stroke and for more
severely disabled subjects. When longer walking distances or
specific speed levels are targeted, as for aerobic training, BWS
can be combined with TW.
Conclusion
Fast walking induces speed-dependent adaptations that improve the overall walking pattern of stroke subjects, with no
observable deleterious effects. On the basis of task specificity, speed-intensive training should reinforce improved walking behavior, while enhancing cardiovascular fitness, muscle
power, motor coordination, and postural control to cope with
the increase in walking speed. BWS during overground
locomotion is preferentially beneficial to low-functioning
subjects, inducing an instantaneous increase in gait speed
while favoring a hip flexor strategy. In contrast with the
conclusions drawn from a recent review on supported treadmill ambulation,30 use of BWS during overground locomotion should be recommended as a useful intervention strategy
6
Stroke
November 2004
in early rehabilitation with severely disabled subjects, as
evidenced by its instantaneous improvement on gait speed,
stepping, weight bearing, and balance.
Acknowledgments
A.L. was the recipient of a postdoctoral fellowship from the
Canadian Institutes of Health Research (CIHR). We are grateful to
Rukshana Gheyara, Laura Galiana, and Roain Bayat for their
assistance, and to the Jewish Rehabilitation Hospital (JRH) Foundation and CIHR for their financial support.
References
1. von Schroeder HP, Coutts RD, Lyden PD, Billings E Jr, Nickel VL. Gait
parameters following stroke: a practical assessment. J Rehabil Res
Dev. 1995;32:25–31.
2. Dean CM, Richards CL, Malouin F. Walking speed over 10 metres
overestimates locomotor capacity after stroke. Clin Rehabil. 2001;15:
415– 421.
3. Olney SJ, Richards CL. Hemiparetic gait following stroke. Part I: characteristics. Gait Posture. 1996;4:136 –148.
4. Said CM, Goldie PA, Patla AE, Sparrow WA, Martin KE. Obstacle
crossing in subjects with stroke. Arch Phys Med Rehabil. 1999;80:
1054 –1059.
5. Richards CL, Malouin F, Wood-Dauphinee S, Williams JI, Bouchard JP,
Brunet D. Task-specific physical therapy for optimization of gait recovery
in acute stroke patients. Arch Phys Med Rehabil. 1993;74:612– 620.
6. Laufer Y, Dickstein R, Chefez Y, Marcovitz E. The effect of treadmill
training on the ambulation of stroke survivors in the early stages of
rehabilitation: a randomized study. J Rehabil Res Dev. 2001;38:69 –78.
7. Hesse S, Bertelt C, Jahnke MT, Schaffrin A, Baake P, Malezic M,
Mauritz KH. Treadmill training with partial body weight support
compared with physiotherapy in nonambulatory hemiparetic patients.
Stroke. 1995;26:976 –981.
8. da Cunha IT Jr, Lim PA, Qureshy H, Henson H, Monga T, Protas EJ. Gait
outcomes after acute stroke rehabilitation with supported treadmill ambulation training: a randomized controlled pilot study. Arch Phys Med
Rehabil. 2002;83:1258 –1265.
9. Visintin M, Barbeau H, Korner-Bitensky N, Mayo NE. A new approach
to retrain gait in stroke patients through body weight support and
treadmill stimulation. Stroke. 1998;29:1122–1128.
10. Dean CM, Richards CL, Malouin F. Task-related circuit training
improves performance of locomotor tasks in chronic stroke: a randomized, controlled pilot trial. Arch Phys Med Rehabil. 2000;81:
409 – 417.
11. Sullivan KJ, Knowlton BJ, Dobkin BH. Step training with body weight
support: effect of treadmill speed and practice paradigms on poststroke
locomotor recovery. Arch Phys Med Rehabil. 2002;83:683– 691.
12. Pohl M, Mehrholz J, Ritschel C, Ruckriem S. Speed-dependent treadmill
training in ambulatory hemiparetic stroke patients: a randomized controlled trial. Stroke. 2002;33:553–558.
13. Murray MP, Spurr GB, Sepic SB, Gardner GM, Mollinger LA. Treadmill
vs floor walking: kinematics, electromyogram, and heart rate. J Appl
Physiol. 1985;59:87–91.
14. Pearce ME, Cunningham DA, Donner AP, Rechnitzer PA, Fullerton GM,
Howard JH. Energy cost of treadmill and floor walking at self-selected
paces. Eur J Appl Physiol Occup Physiol. 1983;52:115–119.
15. Barbeau H, Lamontagne A, Ladouceur M, Mercier I, Fung J. Optimizing
locomotor function with body weight support training and functional
electrical stimulation. In Latash ML, Levin MF, eds. Progress in Motor
Control: Effects of Age, Disorders, and Rehabilitation, vol 3. Windsor,
Canada: Human Kinetics;2003:237–251.
16. Fung J, Barbeau H, Roopchand S. Partial weight support improves force
generation and postural alignment during overground locomotion following stroke. Soc Neurosci Abstr. 1999;25:907.
17. Perry J, Garrett M, Gronley JK, Mulroy SJ. Classification of walking
handicap in the stroke population. Stroke. 1995;26:982–989.
18. Hesse S, Helm B, Krajnik J, Gregoric M, Mauritz KH. Treadmill training
with partial body weight support: influence of body weight release on the
gait of hemiparetic patients. J Neurol Rehabil. 1997;11:15–20.
19. Nadeau S, Gravel D, Arsenault AB, Bourbonnais D. Plantarflexor
weakness as a limiting factor of gait speed in stroke subjects and the
compensating role of hip flexors. Clin Biomech (Bristol, Avon). 1999;14:
125–135.
20. Olney SJ, Griffin MP, McBride ID. Temporal, kinematic, and kinetic
variables related to gait speed in subjects with hemiplegia: a regression
approach. Phys Ther. 1994;74:872– 885.
21. Vaina LM, Cowey A, Eskew RT Jr, LeMay M, Kemper T. Regional
cerebral correlates of global motion perception: evidence from unilateral
cerebral brain damage. Brain. 2001;124:310 –321.
22. Marshall SC, Grinnell D, Heisel B, Newall A, Hunt L. Attentional deficits
in stroke patients: a visual dual task experiment. Arch Phys Med Rehabil.
1997;78:7–12.
23. Murray MP, Mollinger LA, Gardner GM, Sepic SB. Kinematic and EMG
patterns during slow, free, and fast walking. J Orthop Res. 1984;2:
272–280.
24. Hesse S, Werner C, Paul T, Bardeleben A, Chaler J. Influence of walking
speed on lower limb muscle activity and energy consumption during
treadmill walking of hemiparetic patients. Arch Phys Med Rehabil. 2001;
82:1547–1550.
25. Corcoran PJ, Brengelmann GL. Oxygen uptake in normal and handicapped subjects, in relation to speed of walking beside velocity-controlled
cart. Arch Phys Med Rehabil. 1970;51:78 – 87.
26. Wagenaar RC, Beek WJ. Hemiplegic gait: a kinematic analysis using
walking speed as a basis. J Biomech. 1992;25:1007–1015.
27. Cavagna GA, Willems PA, Heglund NC. The role of gravity in human
walking: pendular energy exchange, external work and optimal speed.
J Physiol. 2000;528:657– 668.
28. Ivanenko YP, Grasso R, Macellari V, Lacquaniti F. Control of foot
trajectory in human locomotion: role of ground contact forces in simulated reduced gravity. J Neurophysiol. 2002;87:3070 –3089.
29. Roopchand S, Fung J, Barbeau H. Locomotor training and the effects of
unloading on overground locomotion following stroke. In Columbus F,
ed. Progress in Stroke Research. New York, NY: Nova Science. In press.
30. Moseley AM, Stark A, Cameron ID, Pollock A. Treadmill training and
body weight support for walking after stroke. Cochrane Database Syst
Rev. 2003:CD002840.